A new OS from Google
Okay, fine. One question : what's the telemetry on that thing ?
This is Google, it exists in order to find out everything about you. So how will this new OS feed it data and what data will it feed on ?
A couple of days ago and without fanfare Google went live with Fuchsia.dev, a developer site for its new operating system, currently called the Fuchsia Project. What is Fuchsia? The first the general public knew about it was when a GitHub project appeared in August 2016 (though the repository has since moved to Google's own …
I see that the Google Fanboys are out in force today. sic
Monsieur Monett is asking a perfectly valid question yet he gets downvoted for doing so.
Those of you who did downvote this are more than welcome to go and live in your cozy world where Google and Amazon can satisfy your every need (at a hefty price naturally).
Those of us who benefit from years of experience in the IT world will continute to ask this sort of question especially when the arc slurpers are involved. I also include Microsoft in this.
I deliberately limit my exposure to Google, don't buy anything from Amazon unless there is nowhere else and stopped using anything from Redmond years ago.
Viva la Revolution!
May your life and data be protected from these leeches.
I sure do have a mobile phone (Samsung Galaxy S series) but it is being used exclusively for voice and SMS. I'm also using a credit card from time to time but it is not related to my mobile phone in any way.
So yes, Google might know where I am and what I'm buying (mostly gas and cable subscription) but if they want to throw ads at me, they have to wait until I get home in the evening and power on my desktop computer where gmail runs quite nicely in a different, stand-alone browser separate from my regular Internet browsing activities.
No, you don't get it. Credit card companies/Banks have been using your purchasing data in similar ways to Google (and more invasively) for years. So have phone companies - they know who you are, what sort of person you are, where you've been etc.
All bundled into a nice data profile.
That's possibly because Monsieur Monett's comment implies the sort of rabid Google hate common in these forums.
Don't get me wrong. Google is extremely far from perfect. That being said, the data harvesting is the logical extreme of their core business' goal of giving the best search results and the most accurate ads* that they can.
In the context of this article, it appears that they're trying something where not much progress has been made with a security by design OS. if it's full to the brim with telemetry, that would suck but it's Open Source. The existence of solutions to existing problems around this design is only a good thing and will be out in the open where programmers can see how problems were solved. If it's really good, it may get its own GNU variant (Dahlia is Not Fuchsia).
*The typical response to this statement is that the ads aren't relevant to the user. This is mainly true for people who go out of their way preventing Google from knowing anything about them.
Your apprehensive servant,
If it's a good OS, it'll be used.
Given that most of "us" have an Android phone and or use Chrome as our desktop browser anyway... you have to wonder how much more there is for Google to slurp?
The biggest advantage here seems to be that you can consolidate the slurp to one company XD.
I mean, Microsoft are doing their best to slurp you too.
If it helps protect against malicious apps harvesting your data too, then it's a net positive, and being open source, we might get a telemetry light version too, maybe?
If it's a good OS, it'll be used.
Given Googles current dominance, whether good or bad, if Google decided to push it, it'll be used.
I suspect that plenty of users on here whether overly suspicious of Google, rightfully suspicious of Google, ambivalent, fan or unapologetic hater with a permanent erect finger in their direction to prove it, most of you are united in smart phone ownership
"most of you are united in smart phone ownership"
You must have missed the hordes of self-righteous arseholes who loudly proclaim how they are far too clever to fall for buying a smartphone, pretty much any time there's an article about it.
My favourite from the other day was someone saying that they insist that everyone should contact them through their homebuilt mumble server.
You must have missed the hordes of self-righteous
Not any more than I've failed to notice the hordes proclaiming 'that's it, I'm installing Mint' at every MS gaff, funnily enough, the revolution has failed to happen, probably another AAA title was released in the meantime.
Not many are really willing to swim against the tide as they claim, and it's increasingly like not having a smart phone is seriously hamstringing yourself in the modern world as governments and corps assume you are on the bandwagon, and attempt to herd you onto whatever app they are pedalling in place of an adequate and fit service.
Since you don't mind being slurped, why do you care which company is slurping your personal info ? What difference does it make for you if it's Google, Microsoft, Facebook or the Chinese government ?
As for the telemetry light version, don't count on it even if this OS will be "wink wink nudge nudge" open source.
TBH, anything that gets advertised directly to me goes straight on my 'Do Not Buy' list.
I hate all forms of advertising. I was pretty ambivalent about it until I went to work for an Ad company in 2010. I lasted six months by which time I'd seen a lot of the sneaky tricks they play on us poor suckers.
Since then Ad-blockers have been the norm for me. Searching using StartPage and generally obfuscating my internet footprint has been the norm. VPN's are wonderful for that.
Even so there are times when you have to get into bed with the enemy and use their services. One Time VM's come into play here which are naturally VPN'd out to foil their location sniffers.
Targeted ads are EVIL. Location tracking is NASTY.
Yes, I'm a miserable old bugger but I do value my privacy thank you. Downvote me all you like but it won't change my view of the likes of Google.
Blocking ads is your choice.
In terms of Targeted ads and Location tracking, I can understand why you don't like them, that's also your choice.
Considering that most of the online economy is either ad supported, paywalled, retiail or loses money, it makes sense to try and improve ad targeting because the companies buying the ads also need to get their return.
Hate it as you do, a lot of the web straight up wouldn't exist without ads because the money has to come from somewhere. Do you think El Reg. survives on its merch store?
False dichotomy. It's entirely possible to have sites funded (or partly subsidised) by ad revenue, without collecting vast amounts of tracking data allowing you to be triangulated across 90% of the other sites your visit. Just because we don't live in that world today it doesn't mean it's imposisble. It'd just be different.
The whole targeted ad thing is mostly questionable in it's efficacy anyway.
Oversaturation of advertising certainly is.
I stopped watching TV regualrly in 2012 or so, I find if I am exposed to adverts now, I notice them a hell of a lot more.
Don't get me started on the whole 'we know you bought A1, so we're going to shovel adverts for A1, A2 and A3 thru 1001 for the next couple of months in case you want a second almost identical product one, which may work for short life consumables but not Fridges, cars and sofa sets.
Ads are 100% effective - at presenting items somebody wants to SELL
They are considerably less effective at presenting them to somebody interested in buying.. but the ad-slingers still get paid.
In the never-ending war between (us) and (them), the clear winners are the guys selling the armaments..
That's not actually true.
At least in TV advertising the channel only gets paid for the proportion of the viewers in the demographic that the advertiser are targeting.
This is also true in 'static' advertising. The Register is likely to get significant higher revenue per user if they can show 80% of users are IT professionals and therefore in the target audience. If another advertiser is targeting IT students then they might only pay 10% of the price to cover that 10% of the audience. 100% accurate targeting would allow ElReg to get both sets of revenue at the same time - by showing the 10% of students the adverts for the company wanting to target students, and the 80% of IT professionals the ads from the company only wanting to target IT Professionals. The other 10% might be randomly distributed and they effectively wouldn't get any money for those users
From an end user perspective, targeted advertising simply does not work, for example:-
I booked a hotel for a weekend in London, I chose the hotel in Last Minute but went straight to the hotel's website to book it as it's very often cheaper than the "bargain" sites.
From that point for the next three or four months, on any site that served targeted ads, I received an advertisement for the hotel that I'd just booked, for the same weekend I'd booked.
Now if targeted advertising actually wanted to be useful the ads I'd have seen would be for events and/or offers on in London over the weekend I'd booked. For example, £5 off Tower of London entry or a advertising a food festival in Greenwich (this was on at the time).
Targeted advertising just doesn't work in any useful way.
This isn't a defense or critique of how the targeting works currently, just an explanation of the likely chain of events
Assuming you visited the bargain sites for the hotel in London, which I think you must have done to get the information that booking direct from the hotel was cheaper, then what happened is the bargain sites seeing that you were interested in booking a room at that specific hotel and being unable to see that you did convert on the hotel's own site.
Even pre-GDPR, that information wouldn't be shared between sites. Now it's illegal.
Similarly, the events and offers for around London would come from different sites/organisations, running into the same problem as above.
Yes, I used the bargain site to give me a list of hotels in the area of London I wanted to stay and then did a Google search to find the websites of the ones I was interested in. I compared prices and booked directly with the hotel, as it was around 10% cheaper than the bargain sites for the same stay. Doesn't everyone do this, or do people assume that because the sites claim tom be "bargains", that they are without verifying that?
And there you have it; as I said, targeted advertising does not, nor can it ever, work in any way that is remotely useful unless you act like an unthinking idiot and just book at the first site you come to.
"we see better results from targeted ads than untargeted"
Funnily enough I am reasonably up to date on most things I might want / need to buy, I really don't need some geek's idea of a targeting algorithm to clutter up my reading or viewing etc.
For some reason Google seems to want to restrict my own product searches / comparisons and replace it with theirs I wonder why.
“Half the money I spend on advertising is wasted; the trouble is I don't know which half.” John Wanamaker (1838-1922)
You are entirely correct that advertising has been around forever.
Put a magazine ad out, you choose the magazine for its demographic - you'll do better getting an ad for an upcoming racing game into Nintendo Power than into Playboy. However a lot of the readers wouldn't be interested in the genre, wasting a proportion of your marketing spend by putting an ad in front of somebody who doesn't care.
What PPC and Display do, and what Google wants this data for, is to try and figure out who is interested specifically in racing games. Not only that, but you can see how many clicked an ad and how many converted to get the most out of the advertising budget.
Disclaimer: a lot of the following paragraph is Google's goal and is currently far from perfected
This is why you can always trust Google in the way you would trust an enemy - they will soldier on to the goal of knowing everything about everybody. If they know you, they know what you mean when you search for something. Their SEO results are the driving force here, if those were bad then nobody would use their search engine. Because their SEO results are good, they get a lot of eyeballs on SERPs and can combine your search with what they know about you to present ads that are relevant. With the tracking on sites, the advertiser/retailer can see what is working and what isn't working to a degree that would be unthinkable 15-20 years ago.
As has been pointed out, I am indeed one of these marketing bottom-feeders. Take that as you will, but even when working on the accounts, I didn't care who you were. If people searching for a keyword were converting, then we paid more for that keyword. If a demographic were converting well, then we would bid more for that demographic. Trying to figure anything out about an individual was (a) impossible and (b) so far from worthwhile as to be laughable.
My final point on the subject is that a lot of the common complaints about ads are not Google's fault, they are the advertiser's. It's the advertiser who decides that remarketing fridges to people who have already bought a fridge might want to see more fridges in ads. There are also limitations in what you can see, Go to Curry's site looking for a Fridge but then convert from Argos, then Currys will probably remarket fridges to you because they haven't seen that you've converted already.
One Time VM's come into play here which are naturally VPN'd out to foil their location sniffers.
The problem is it is very difficult to hide yourself from those that want to find you. Everyone has a routine, and even if you stop ads, and keep deleting your cookies (as I do) it is still difficult to hide. Think of the websites you use every day. It only takes a handful of those to either uniquely identify you, or put you in a very small set of people who do likewise.
If they have web log entries for your sites they can piece together that it's you. google-analytics.com is used by nearly everyone, and the jquery code you need for a website is often served from some cdn such as ajax.googleapis.com. There, they have you!
Hmm, your connections only ever occur during UK hours. You always read English language pages. You like to look at a lot of sports pages. It doesn't take much to narrow you down to a small set of people.
First of all, AI advertising technology is extremely accurate. AI advertising looks for potential buyers into AI database; that is, almost 100% accuracy of advertising targeting is guaranteed! and the probability of a subsequent purchase can become almost absolute! because AI advertising technology compares many thousands of patterns, while now only very few are compared; compares with respect to the importance of patterns (which does not exist now and cannot exist without AI-parsing).
Secondly, AI advertising technology is looking for a potential buyer regardless of his activity on-line - it's aimed at the patterns only, and not at his activity which caused the arrival of them. AI does not need to spy and steal!
Thirdly, such attention to the patterns can bring us all absolute anonymity on-line - advertisers do not need to know who owns what profile, they do not need to know our preferences, what we saw and read - they need only our patterns.
However, I can refer to Microsoft's success, cannot I? "Microsoft has significantly improved the MT-DNN approach to NLU, and finally surpassed the estimate for human performance on the overall score on GLUE (87.6 vs. 87.1) on June 6, 2019." That is, Microsoft proved my patented method for what I called "lecture noise" purging, which used to be the main practical obstacle to the practical existence of this database; where "lexical noise is typically superfluous predictive definitions that do not explain the central themes contained within the digital textual information and, accordingly, removal of such noise often results in an improvement in the quality of the structured data."
So despite the fact that I could not produce and demonstrate my AI database yet - all its theoretical justification proved by third and completely independent parties. Which gives me the right to talk about AI database almost as a given in perception, right?
Thus, the methods and system of advertising have changed radically, I changed them once and for all.
They have a consumer robot division.
My guess is they'd love to eventually have a Streetview-like slurper going round the inside of people's houses, identifying items to allow better targeted advertising, and sold to people like it's some kind of robot butler to make life easier for them.
When google knows where you left your keys, that bolt you couldn't figure where it came from, your baby, or that last piece of your jigsaw puzzle, everyone's going to be signed up in an instant. (almost added 'your phone' as a popular misplaced item, before checking myself).
When google knows where you left ...your baby... everyone's going to be signed up in an instant.
I know where I left him, but which cupboard is he in now.
Anyway, no use to me; I've already got parent sonar.
"What are you doing?"
"Nothing!" (how do they do that?)
"the data harvesting is the logical extreme of their core business' goal of giving the best search results and the most accurate ads* that they can."
Their core business goal is to sell ads, nothing more. Literally everything else they do is in service of their ad business. That rampant spying is the logical extreme of that goal in no way makes makes it OK that Google does that.
Source code published? Then it won't be in the kernel.
Looking at this, it seems to be a next iteration of what Unix was supposed to be. From a CS point of view it sounds very interesting. Security built in at the object layer. Graphics and filesystems as easily replaceable components. All of that would make covert telemetry by the OS far more difficult, not easier.
The telemetry is going to be, as usual, by the non-OS blobs that actually make useful products. Separating them out in this way probably makes it easier, not more difficult.
It may have been announced now to stop Huawei's Hongmeng dead in its tracks since maintaining Android compatibility with apps would be a dead end.
I'm pretty sure if Google wanted to publish a modular kernel with the telemetry jigsaw piece missing, that would be possible.
As you can see, I trust them as far as I can throw them.
I don't see why Hongmeng would be dead in its tracks because of this. Fuchsia suffers from the same problem as Android does which is Google (or the POTUS via Google) can pull the plug on non-US corporations that use Android any time they want.
Especially since this looks like a way to move away from Linux and GPL - which would allow them to publish only what they think suits best their business model - have willingly serfs working on the bits that don't bring you much revenues, and concentrate on how to slurp every bit of information out of users.
"Yep, Wikipedia says BSD, MIT, Apache 2.0 licences apply but no GPL"
...i.e. the more restrictive licences. I'd be willing to bet that element is going to put off quite a few developers from helping out on this project which will ultimately make Google more wealthy and more dominant but which won't ultimately be openly shared. **** off, Google.
News just in: the GPL means having to ask the lawyers again and again. This is why the GPL lost.
If they have proprietary stuff that they develop then they don't have to make it open source no matter what. MySQL did this with its own commercial add-ons to the "database" of that name.
"Source code published? Then it won't be in the kernel."
If it's not in the publishes source code, then it won't be in the kernel if you've built it yourself from that code. But I see no reason to assume that the published source code is the only code that's running in a binary that I've been handed.
The differences can be substantial, and unless you're decompiling the binaries, you have no real indication of what the differences actually represent. They could be the inclusion of telemetry code, they could be the result of using a different library or version of a library, or they could even be the result of different optimization methods.
A telemetry package would not be likely to stand out, particularly since telemetry code tends to be reasonably small and simple.
There was a "next iteration of what UNIX was supposed to be". It was called Plan 9, and was waaaaay ahead of it's time!
I think AT&T/Lucent/Alcatel either open-sourced it, or put it under a permissive license. There's an x86 build out there somewhere.
Why on earth would they want to put telemetry into the kernel? I'd expect to see it in some kind of locked-down hardware that they control completely and is required for certain functions, say hardware acceleration. Wasn't there something along these lines in one of their phones?
But even then, do they really need more data than they're already getting? How many people switch browsers, phones, etc. because they're really worried about privacy?
It's a microkernel architecture. The bloody FILESYSTEMS don't run in kernel space. No doubt if or when they commercialise it, they'll use it for the typical Googly evil ffuts, but it won't be in the kernel.
And as TFA points out you can download the source and have a look if you care to -- just like you can DL the open source Android kernel and look there for evil tracking gadgets.
"And as TFA points out you can download the source and have a look if you care to -- just like you can DL the open source Android kernel and look there for evil tracking gadgets."
You can read all the source in the world ... but are you certain the toolchain that turns the source into code the computer can use is trustworthy? See: Reflections on Trusting Trust.
And then there are little gems like the Intel Management Engine ... Do you really think that
Bruce Schneier has a counter to your "Trusting Trust" article: Countering "Trusting Trust".
As for the hardware route, what if you produce a version for hardware you can vouch? Remember, it'll be open source, and if Alphabet try to hide bits, the compiled results won't match, raising a red flag.
Of course there is. I have to write a C compiler that's capable of handling every aspect of modern C because somewhere in its massive codebase, Google has definitely used all the things you never think about, oh and also I'll need a C++ compiler too while I'm at it. I also have to write a compiler for dart, go, and rust. However, I'm not worried that they're really compromising the toolchain. I don't actually need the compiled to differ from the source to be worried (though I think it will happen).
First, there will be a bunch of blobs that need to be added to the kernel to get it to do anything. Any or all of these might harbor any malicious functionality, just like now. There is not a good way to avoid that. I'm sure critical functionality will not be available in the open components, and Google will have nicely built all of that in a closed-source component. After a few years, someone will build an open source replacement for it that kind of works a little bit on some apps but you'll have to compile it yourself, root the device, and do some assorted hacking to actually replace it and also it will break a lot. In addition, without the requirement from GPL to release any changes as open source, manufacturers and mobile providers are free to do the same thing to the kernel that they have been doing to the layers above it. Can I say no thanks?
I don't think Fuchsia will be much worse than Android in the sense that consumer devices will contain a similar amount of spyware and irritating or potentially unwanted bits, it will be difficult to impossible to remove or even disable them depending on device model, and very few people would even try. However, given the choice, I would see Android as much better because we already have years of experience getting around some of this. We have Lineage OS, which, for all its flaws and limited device support, is a trustworthy OS that can actually run on a relatively large assortment of devices. At best, Fuchsia means a return to square one to do all this again. But it could be far, far worse.
Lockheimer used to work for Be and a lot of this stuff sounds a lot like the BeOS.
IIRC one of the problems with wanting to put stuff in userland for security reasons was that performance on x86 was shit due to the overhead of context switching. This not necessarily the case on the different CPU architectures. BeOS, now reborn in open source as Haiku, was always a dream to program for with a clean and extensive OO-API combined with the microkernel allowed for a very responsive UI, low memory requirements and hardware acceleation.
What's not to like?
@Peter Ford: Damn, you beat me to it on the "fucks-ya" pronunciation. That was the first thought I had about it!
(But, yes, an interesting step for GRUgle to name an OS using a word that is usually somewhat mangled (or misspelt: fuschia) when pronounced by a native speaker of English...)
one of the least trustworthy companies on the planet
Really? I can think of hundreds of companies (starting with Big Oil, Big Tobacco, Big Pharma, etc.) that are a lot less trustworthy. Yes, Google does collect and, horrible world, monetise personal user data. (And who doesn't? — I've just been reviewing a proposal for some pretty invasive user-tracking). But, in order to abe to collect the data. it has to produce products and services that people want to use and this means getting at least some of the stuff right.
Google's engagement with open source is also interesting, in that it is often proactice in making source available, as with Fuchsia. While this probably stems from a position of enlightened self-interest it's still a damn sight more open than most tech companies.
But by reducing the debate to goodies versis baddies, you effectively exclude yourself from it. If you are worried about privacy and data protection, and we should all be, then pushing for the effective enforcement of good legislation is a better use of your time.
I'm not a google fan. I think they are no different to any big company in that when they get to a certain size, they only see reflected ideals, not those form outside their ecosystem. I don't trust them.
That being said, I am actually excited about this project. I have long held a belief that we need a new OS philosophy. Linux, whereas my favourite, I think is not moving in a direction I care for and I think that where will it be in 10 years, yet alone 20. BeOS I wanted to see flourish.
Regardless of any motives, I think that there are a *lot* of bright people in Google who really care for the same things in a new OS and that regardless of the business as a whole, will have the heft and resources to push a new OS out into more mainstream.
So I am looking forward to it.
Google's engagement with open source is also interesting, in that it is often proactice in making source available, as with Fuchsia. While this probably stems from a position of enlightened self-interest it's still a damn sight more open than most tech companies.
If you can't see it's Android/Chrome/gmail all over again you're blind. First Google bring in outside help, developer support, mindshare, and whathaveyou when they need it at the start, once it takes off the closed source binary blob will appear and protocols and APIs will suddenly get more closed.
I'm not batting for Google but I don't agree. Seeing as I've been using AOSP on my phones for several years, don't use Chrome, except for testing and have never used GMail. I think Google's use open source is interesting. There is certainly an argument to be had for the secret source approach, which is certainly true with Android and with stuff in the data centres. But Google does release stuff to the world early and does, in my experience, generally engage with developers more seriously than other companies do. I think they understand that making some stuff free and open source can effectively stifle the competition while suiting the service-based model. It can also encourage adoptive addiction (TensorFlow).
But they do engage constructively in a large number of projects. Even with the current shitstorm about Chrome extensions, their stewardship of Chromium is far more open than, say, Oracle's of Java. Their work on WHATWG and things like Polymer have definitely improved things for web developers and users. As has their work on video codecs (for which of course Widevine is the locked down, secret source). Enlightened self-interest runs through all of this but doesn't mean the projects aren't interesting or useful or always require their permission. And their is merit in this OS.
Big Oil pursue profits, and destroy countries in their wake.
Big Tobacco pursue profits, and destroy lives to do so.
Big Pharma pursue profits, and destroy personal economies to do so.
Big Google are actively seeking to exert societal change across the world, and are doing their best to destroy civilisation to do so.
I class Google's activities are far more evil than the others you've mentioned, and I don't trust them or their motives in the slightest.
(Example: After the BBC mentioned that Reddit had "quarantined" a large forum I used Google to try and find it. No links to the forum on the first three pages of search results, because Google don't want you to go there. Bing's response to the same search query was a link to the forum as their top search result)
I would expect Bing to comply with robots.txt too.
Interestingly going to https://www.reddit.com/robots.txt returns 'Not found' - looks like they're providing a different response based on who's making the request (user agent or IP address range?)
Based on https://canicrawl.com/reddit.com/ they're not naming any specific forums.
"doing their best to destroy civilisation"
"Google don't want you to go there"
The page has a noindex,nofollow tag. Reddit have asked search engines not to index it or follow links in it. Google are respecting site owner's requests. Bing, is not.
"I can think of hundreds of companies (starting with Big Oil, Big Tobacco, Big Pharma, etc.) that are a lot less trustworthy"
Not "less trustworthy", rather "equally untrustworthy".
"pushing for the effective enforcement of good legislation is a better use of your time."
What makes you think I'm not?
"IIRC one of the problems with wanting to put stuff in userland for security reasons was that performance on x86 was shit due to the overhead of context switching."
I believe ARM is no different in this regard because the architecture doesn't include hardware features for this (which would be required to avoid the associated penalties of switching back and forth between kernel mode and user mode). This is especially true for parts of the hardware that historically needed close-to-the-metal coding for performance reasons such as graphics and networking (both of which are latency-sensitive, recall the original Windows NT).
The issue with x86 is which code can access the I/O instructions and physical memory directly - most hardware usually use a combination of the two - being mapped to some I/O ports and physical addresses.
You can move both accesses up to userland, but then you lose some hardware security - the CPU itself won't allow user code to run those instruction or access that memory - and you need to deploy other security mechanism.
Anyway, even with multiple cores the issue is when a thread needs a ring transition - the required checks and call setting up require CPU cycles (even if they've been reduced by the SYSCALL/SYSENTER opcodes) - execution might not be passed to another thread on another core.
Your model would be message based - split the CPU between cores executing kernel and core executing user code - and pass messages between them. It could be less versatile if you don't know workloads in advance.
Thanks for the note, though my information (from someone who went on to work at ARM) was that x86 was a lot slower than other archictectures. Be did move the networking code into the kernel for performance reasons and Haiku has it there as well.
I'm trying to think of the first OS that touted capabilities…
Ken Thompson demonstrated that in his infamous "Reflections on Trusting Trust" speech at the 1983 Turing awards. He built a version of cc, the c compiler, that had enough logic to recognize when it was compiling the Unix login code or when it was compiling the c-compiler. When it detected that the compiler was building the login code, it would inject itself into the resulting binary. If it detected you were compiling the Unix login code, it would insert an extra line into the case statement that processes the username. In this case, if you used the username 'ken', the case statement would just jump over the code to check /etc/passwd and just set the UID/GID to 0 and proceed to setting up the session.
You could have clean source for both the login page and cc, then re-compile the compiler before building Unix and still have his backdoor present on your system. You could only get rid of it by using a different compiler, used an old compiler to build the new one, or manually remove the backdoor code from the binary.
And Bruce Schneider produced a counter to the exercise using multiple compilers running against each other to make an evil compiler trip up, and it's possible to build a clean compiler in steps going all the way back to a hand-assembled program too simple and overt to subvert.
Why run the GPU driver in userland? It is effectively the decision never to let the OS kernel take advantage of GPU hardware. Seems bonkers to me, but what do I know?
I guess Fucksya (oh the joys of dyslexia) will end up as free and open as its API Council - every member of which has a google email address.
Am I just an old cynic or is posting the code open source just an invitation to enter the jaws of the ecosystem and presently be bitten by license fees?
Modern GPUs are very complex beasts which require huge drivers which cannot be effectively audited. Running all that code in kernel mode is asking for trouble (security & stability). Windows since Vista switched to back to mainly user-mode GPU drivers (with a small portion inside the kernel). I've had the NVIDIA driver crash before, the screen goes blank for a second and then a pop-up notification appears telling you the driver crashed.
It's a microkernel OS, the whole idea is that more or less /everything/ runs in userland drivers included, so it's no surprise the graphics driver does too. That's fine; it's theoretically better in terms of reliability etc, historically the problem with microkernel OSes has been poor performance because of all the extra context switches and the added overhead of RPC calls between driver processes instead of just jumping to another function inside the kernel.
That said, as noted, virtually all of the GPU driver is in userspace on all modern OSes anyway; the kernel bit is just enough to transfer the data and gpu code into the gpu and kick it off. Why on earth would you want e.g. the shader compiler or the OpenGL API in the kernel when it doesn't need to be?
Yes of course - since microkernel gets out of the way entirely, you can have better performance because that removes the context switch. This is called kernel bypass and it can significantly reduce latency (and improve throughput with it). That's where Intel's DPDK gets its performance boost from and also why Solarflare is so liked by the financial industry.
In principle yes. You need a light weight comms protocol that doesn't copy buffers of information around as context switching occurs. ICL in the nineties had a system (Goldrush) that used a mircokernel and had a very light weight inter node comms protocol that didn't copy buffers on context switching. For it's time, it was very fast. It's the sort of thing you can do if you control the whole stack from hardware, up through the kernel and into user land.
In principle, yes, but in reality nVidia and AMD are in hot competition and keep a lot of Trade Secret Sauce around because of that. You see the same thing with IoT/mobile SoC manufacturers who release their drivers only in blobs for fear of Giving Information to the Enemy.
IOW, sometimes there are bigger concerns.
Sounds to me much more like the AMOEBA (pdf) operating system which, I believe, introduced capabilities as an addressing/security addressing method to operating systems. It also spawned the FLIP (pdf) protocol as way of communicating efficiently between capability based entities on a network.
Andrew S Tanenbaum has a lot to answer for. Probably in a good way.
The concept of capability-based access control well pre-dates the Amoeba system, see the following links containing some examples dating back to the 70's:
(Mr. Hardy's link collection lists a number of systems that are not strictly speaking capability-based, see his notes on the individual links.)
Also, since this is a British forum, it might be appropriate to especially point out the Plessey System 250 (which was not just an OS but a capability-based hardware architecture with an OS to match):
"The project may be Google's future operating system for all kinds of devices – this actually looks likely – or it may be scrapped and never used in anger."
It's a Google project. There's approximately a 100% chance it will be scrapped within a couple of years regardless of whether it's their future operating system or not. The only question is whether it will be allowed out of perpetual beta status first.
God help me, I have that, along with everything else between Witchwood and Burning For You... what're the odds?
Pretty dark lyric though:
"A child denied all love can't weep
But bravely bears her life alone
So Fuschia as you try to sleep
You dream of friends you've never known [...] "
This is basically Google's play to move their spyware directly into the operating system kernel (or, since it's a microkernel, a "userspace module" ... hey, how's HURD doing?). But they'll abandon it once they have their own chip fab and can move their spyware directly onto the silicon.
The orange swamp thing has recently shown how vulnerable the economic use of foreign technology can be. Imagine the Trump admin telling Microsoft to stop doing business with China. I was in China in 2008 when Microsoft killed all the pirated XP machines.. That was interesting. Me and my colleagues were the only ones with working XP machines on campus. The disruption of Huawei and causing it such damage is due to the vulnerability of the rights to use being related to foreign law. In a way exactly the game China has played with foreign companies for quite some time, also basing itself on national security. But the idea that China will open its borders to Western tech companies and an end to censorship which was rumored to be the case is pretty deluded. 90 Million Chinese depend on that system for their privileges. They will stick it out till 2020 and if need be longer.
One thing that open source has to offer here is the lack of political sources of disruption. Considering the increase of disruptors and disruptions worldwide, it may well be the obvious economic choice for the industry.
Open source has nothing to do with end-user digital freedoms. Besides, in case you haven't noticed, no digital freedoms can thrive on locked-down proprietary hardware.
Don't worry, measures preventing access to hardware will make sure disruptors will safely laugh their way to the bank. And all these measures will be open sourced of course.
Or open source is empowering authoritarian states with surveillance and censorship tools they would not have been able to build themselves?
We believed the internet would have brought mode freedom, and instead it's being used to reduce it - there's no assurance open source code will be used for good only.
It sounds like one of the primary design features is to make it easier for Google itself to update the device -- something that I don't trust Google to be doing.
Maybe, but many phone carriers or manufacturers can't be trusted to be bothered to update something you've already paid for a couple of months ago either, not when you might be encouraged to pay again.
This really is another case of having no friends. The G has tried desperately (& failed) to get phone manufacturers & operators to issue critical security updates. From a security standpoint, you might consider cell phones to be the first IoT devices.
And since we know that 99.9% of users won't do the updates themselves without at least a push notification from _somebody_, what do you do?
It's too early to get this depressed.
There are two possibilities here. One, we're seeing the beginning of something huge that will eventually replace both Android and Chrome OS. Or two, we're seeing Google fling yet another piece of mud at the wall to see if it sticks yet they'll get bored of it in a few years, withdraw it and send out a 'thanks for playing, bye!' message to all the people who have heavily invested in it.
Given Google's behavior of the past few years, I know which one my money's on.
"Given Google's behavior of the past few years, I know which one my money's on"
^ Absolutely this. Google's now even surpassed Apple as the King of Abandonware with discarded projects littered all over the place and Fuchsia may yet end up in that pile of rejects too.
Picasa was used and appreciated by millions...so it was killed off. The Nik filter collection was bought and then sold on and if I recall correctly one of the robotics concerns acquired by Google has now been terminated too.
Finally, memo to the obvious Google employee who keeps on voting critical comments down:
1. Get a life;
2. Put the entire Fuchsia project under a GPL licence and watch magical things happen (like a great increase in developer interest).
They've never developed an entire OS.
Android was bought in and deliberately used Linux kernel and a form of Desktop Java to leverage the Mobile Java developers on S60 etc. Sun wouldn't licence full Java for Mobile, at least least not free like desktop.
The Android GUI has little changed since Google bought it and the OS is still like a work in progress.
ChromeOS is the attempt to use only Cloud/Browser on top of Linux.
Also will it suit users and purely local processing or somehow be crippled without always on Internet, Cloud and a Google Account with Google tracking.
I'm sceptical. What have they ever developed apart from search?
Look at their attitude to law and privacy, stuff always in beta or abandoned.
Do they make MS look trustworthy?
I know very little about OS design, but I find it very interesting to see somebody attempting to create one from scratch, "The Right Way", with the depth of knowledge to understand long-term requirements and common pitfalls, and with enough power and resources to make it succeed.
Essentially, this could potentially have as much influence on the history of computers as the creation of Linux.
For sure, the advent of photography changed the way we perceive imagery. The view through the lens -- monocular, limited, incredibly detailed -- became a standard of "reality". The still camera, the movie camera, the television camera, all sang us a song that got inside our heads. The song began to tell us it was really how we see things. Photographs, movies, TV, YouTube: we think they are "real", somehow.
Surveillance and data expand that fake reality. Advertisers look to customize propaganda to fit consumer profiles. This is a quote from "a data scientist" involved in marketing, anonymously referenced in an article in the Guardian: "We can engineer the context around a particular behaviour and force change that way. We are learning how to write the music, and then we let the music make them dance."
The "them" in the sentence is, of course, us.
And so, to Fuschia: if deep surveillance can be inserted in an OS (as Microsoft is doing) then it's even more effective (and invasive) than putting it in browsers and search engines. It's almost like having a spy in your own ... oh, wait, Alexa. Right, we're there already on that count.
But Fuscia is supposedly open source. Can it be forked, cleanly? Perhaps a good strategy for Google is to build a clean, fast, secure OS and then engineer the ecosystem for surveillance rather than embed it in the kernel. Hell, I know nothing about OS design. Tell me if this is all off-base.
Sure, just like Android. How far any fork has gone?
Moreover once they are no longer bound by GPL and someone else's copyright on code, they will have much more freedom on what they open source and what they don't.
It's the Apple model - only parts of macOS are open source, those parts they think its advantageous for them to open source, while keeping closed all those parts which they deem dangerous to their business to open.
Make those parts essential for any usable system to work, and remunerate developers who write apps for your system.
Rather then purchased it, it is obviously going to be one of those Google projects that stays in beta for a few years, then gets axed with little fanfare. The only thing Google invented themselves that survived in the long term is search. All the rest of their long lived offerings like advertising, Android, Maps, etc. they bought.
"The only thing Google invented themselves that survived in the long term is search."
Not true at all. Not even if you squint. I had a home-grown search function on my portal before google existed; so did many other portals. Prior to the Web, Knowbot was late '80s, Archie in 1990 ... whois accesses a search engine, and has from the year dot (late 1970s, standardized in the early '80s). Etc.
More commonly known today, AltaVista, InfoSeak, MetaCrawler, AskJeeves and Lycos all existed before google.
Yes, I was talking about Google inventing their own search, rather than buying the tech elsewhere as they have with all their other successful and/or long lasting products.
It is pointless to argue about who created a product category. Who invented the "PC", was it IBM, was it Apple, was it Altair? Altair brought it to the ultra geek hobbyist, then Apple brought it to the mass consumer market, then finally IBM (and Microsoft) brought it to the world. Regardless of other "search" applications, Google is the one who brought it to the world to the extent that it became a verb. Nobody ever said "I'll have to altavista that", let alone "I'll have to ask Jeeves".
You're missing the point. "Google invented search" means that Google invented an algorithm or rather a set of algorithms they used to create a search engine that was better than the others at the time and is still good today. Of course they didn't invent the concept of searching resources. Similarly, Gutenberg invented a useful form of printing press, but didn't invent the concept of printing or the printing press as a type of product. Arguing that he gets the credit for stuff that existed before him would be weird, but so would be attempting to deny him the credit for developing a technology that proved to be a very successful and influential implementation.
Google's innovation was page ranking, which is a bit like academic citations for the Internet. Since they started they have spent a lot of time trying to stop people gaming the system.
Everything else they bought or developed in-house and then dropped.
Fuchsia could be interesting, but might be a little bit evil.
So permissions exist, nice; will it also explicitly _require_ that applications continue to work seamlessly within their granted permission envelope regardless of exactly which combination of permissions gets granted...? Because I can already do "you can try to refuse but then I'll crash or outright refuse to install" already even on the oldest of Androids still in use. And if it does that, will there be options to not just permanently grant a permission but also options like "just this once and absolutely never, ever again, don't even try asking"...? How about "just this once, but you may ask again"...? Or "granting permanently, but I want to be notified / see it logged every single time you _do_ use this"...?
How about some granularity like "yes you can read and write my SD card / non-system storage, but only concerning your own stuff, as opposed to a blanket permission to index or snatch and upload any file I might have on there"? Or how about "No you absolutely may NOT access my contact list for your free unfettered perusal - if you offer to do something with a contact, ask the system to prompt me to select ONE specific record and let you have its details, assuming I accepted your request for contact info access at all in the first place"? And how about "no you may absolutely NOT read anything concerning my contacts - what you may do is attach your own data to them if I allowed that, as hosted by the system, and presented to me by a system dialog when needed at your request, while you only ever see a bunch of GUIDs and the data you attached to them"...?
Because frankly, anything less is worthless security theatre at best - obtaining permission is not a "formality", you either _are_ in control or you _aren't_. And I think I know the answer to all of the above...
"Because frankly, anything less is worthless security theatre at best - obtaining permission is not a "formality", you either _are_ in control or you _aren't_. And I think I know the answer to all of the above..."
So how do you teach Joe Stupid all this when changing the channel is a challenge for him?
As in, for all you ask, you still (like in the comic) have to deal with Dave.
I'm convinced that advertising is a scam. Does it work sometimes? Sure, for good campaigns (Old Spice's "The man your man could smell like" campaign was wildly successful), however for the vast majority of companies I have a feeling its not worth the spend; nowhere near enough bang for your buck. Correct me if I'm wrong, but didn't Facebook come under fire for vastly over-stating the value of advertising on it?
Facebook were generally overreporting the total views or total traffic. The traffic numbers they reported indicated that there was a lot more space for extra advertising to fill rather than the value of the advertising on the platform.
One of the clients I've worked for averaged out at 10% COS (£1 advertising -> £10 revenue) for CPC.
It's not about the grand campaigns (Old spice, Dear Kitten), a lot of advertising online is about getting the ad under the nose of somebody who is about to buy and ensuring they click on your blue link.
Biting the hand that feeds IT © 1998–2021